Official code of our work, AVATAR: A Parallel Corpus for Java-Python Program Translation.

French-English translator using word embeddings, bi-directional encoder, and decoder with attention

AdonisJS Lucid internationalization models.

Build a deep neural network with keras that functions as part of an end-to-end machine translation pipeline

Coding education project - Speech Process & NMT

🤖 ➡︎ 🧑🏻‍🦰 Human-Like Story Generation from Caption Using Seq2Seq Model

基于CSANMT模型的翻译服务,支持多语言互译;tensorflow,nvidia-gpu推理应用; 使用了python的fastapi框架. 小而美的翻译应用

By clicking on the word see the translation

Assyrian translation model using huggingface

Text file language translation app that translates .txt, .csv, and .xlsx files on the local device using lightweight, open-source LLMs from Hugging Face (no internet connection needed). Uses batch processing and token chunking which provides the ability to translate large file sizes while preserving semantic meaning across paragraphs / sections.

Testing harness is an integrated dashboard to evaluate the capabilities of various LLM models

Experiments on Neural Language Embeddings

Use Encoder Decoder architecture with LSTM for English->German translation network

Classic SMT and feed-forward neural networks: LT2212 V19 Assignment 4 group Rho

The Backend service for Testing harness. Testing harness is an integrated dashboard to evaluate the capabilities of various LLM models Resources

This project aims to build a Transformer from scratch and create a basic translation system from Arabic to English.

LingoLite is a lightweight, mobile-optimized neural machine translation (NMT) framework designed for efficient multilingual translation on resource-constrained devices. Built with PyTorch, it features a modern transformer architecture with state-of-the-art optimizations for mobile deployment.

🅰️⇄🈂️ A LSTM/GRU-based machine translation modeling and evaluation pipeline in PyTorch